Cap Tables & Dilution

What a cap table actually is

A capitalization table tracks who owns what slice of a company. Every share, every option, every convertible note, every SAFE. In its simplest form it's a spreadsheet: rows are shareholders, columns show share counts and percentages. The complication is that there are different ways to count shares.

Issued and outstanding: the actual shares that have been issued to actual people. Founders' common stock, granted options that have been exercised, preferred stock issued to investors.

Fully diluted: issued and outstanding plus the unissued option pool, plus anything that could convert into shares (warrants, SAFEs, convertible notes). This is the number investors care about during diligence, because it's the worst-case denominator.

When a founder says "I own 60% of the company" without specifying which basis, they almost always mean issued-and-outstanding. When you, as the investor, run dilution math, you almost always want fully-diluted.

A worked example: incorporation through Series A

Let's track a fictional startup. Two founders, Alice and Bob, split equally. They incorporate with 8,000,000 shares of common stock between them and reserve 2,000,000 shares for an employee option pool. Call it 20% of the fully-diluted total at start.

Day one cap table (10M fully-diluted):

They raise a $2M seed round at a $10M post-money valuation. The seed investor gets 20% of the company in newly-issued preferred stock. The option pool stays at 20% post-round.

To preserve a 20% option pool after the round, the existing 20% pool needs to be expanded pre-money. This is the "option pool shuffle" and it's covered below.

Post-seed cap table:

The founders went from 40% each to 32% each. Notice they still own the same number of shares (4M each). The pie just got bigger. That is dilution. Your stake stays the same in absolute terms; it just represents a smaller slice of a now-larger company.

Two years later they raise a $10M Series A at a $40M post-money. The Series A investor gets 25%. The option pool gets topped up to roughly 15% post-round.

The full math is straightforward but tedious; what matters is the trajectory. Founders typically land in the low-20% range each post-Series A, under 15% by Series B-C, and frequently under 10% by IPO or large exit. That isn't a bug, it's the structure.

The option pool shuffle

In the seed example, the option pool got expanded before the new investor's money came in. That increase came out of the founders' shares, not the investor's. From the investor's perspective: they want enough headroom in the option pool to hire the team to grow the company they're now investing in. Reasonable. From the founder's perspective: the larger the pool, the more dilution they personally absorb pre-financing.

Brad Feld has written about this for two decades. His advice: build the pool from a real hiring plan, not a default 10% or 15%. If the company genuinely needs to hire a CTO, a VP Sales, and eight engineers in the next 18 months, model what those grants will look like and size the pool accordingly. Anything beyond that just cheaply transfers ownership from founders to existing equity.

Liquidation preferences: the pref stack

This is where it gets dense. Liquidation preference is the term that decides who gets paid (and how much) before anyone else, in the event of a sale, merger, or wind-down.

The default term in modern early-stage deals is 1x non-participating. The investor gets the greater of:

• their original investment back (1x what they put in), or
• their pro-rata share of the proceeds as if they'd converted to common stock

That's it. Pick whichever yields more. They don't get both.

Participating preferred (the "double dip")

Some deals use participating preferred. Investor gets their preference and their pro-rata share of remaining proceeds. Both. As Scott Kupor puts it in Secrets of Sand Hill Road, this is "where the VCs get to double dip. They get their money back, plus they get to participate as an equity investor."

Participating prefs can be uncapped (worst case for founders) or capped at a multiple, e.g., capped at 2x or 3x of the original investment, after which the investor converts to common and stops participating.

In modern early-stage venture, fully participating non-capped prefs are uncommon. Kupor notes that severe anti-dilution and participating prefs are "kind of few and far between" in the current market. If you see it in a Series A or earlier, ask why.

Multiples (2x, 3x)

Standard is 1x. 2x or 3x preference means the investor gets 2x or 3x their money back before anyone else sees a dollar. These show up in distressed rounds, late-stage rescue deals, or investor-friendly markets. They are red flags in normal circumstances, especially at the early stage.

Seniority

When there are multiple rounds of preferred, the order in which preferences are paid matters.

• Pari passu (same priority): all preferred gets paid back proportionally. Most common today.
• Strict seniority: later (Series C) gets paid before earlier (Series A). Aggressive but used in some deals.

The full set of preferences across all rounds is called the pref stack. Modeling an exit waterfall requires walking the pref stack in order.

A worked exit waterfall

Same fictional startup. After Series A, the cap table looks like this (simplified — option pool is fully granted into employee equity):

The cap table alone doesn't tell you who walks away with what. The pref stack does. Below: same cap table, same $50M exit, four different deal structures for the Series A. Click each to expand.

▸Scenario A — $50M exit, 1x non-participating (modern default)

Holder	Math	Payout
Series Seed	max($2M, 15% × $50M)	$7.5M
Series A	max($10M, 25% × $50M)	$12.5M
Founders + employees	60% × $50M	$30M

Both prefs convert to common (each chose the larger conversion number). Founders take $30M.

▸Scenario B — $50M exit, 1x participating uncapped

Series A takes its $10M off the top, then participates pro-rata in the remaining $40M.

Holder	Math	Payout
Series A	$10M (pref) + 25% × ($50M − $10M)	$20M
Series Seed	max($2M, 15% × $40M)	$6M
Founders + employees	60% × $40M	$24M

The single word "participating" moves $6M from founders to the Series A. Series A's IRR jumps; founders' take drops 20%.

▸Scenario C — $50M exit, 2x non-participating

Same conversion math, but the preference doubles to $20M.

Holder	Math	Payout
Series A	max($20M, 25% × $50M)	$20M
Series Seed	max($2M, 15% × ($50M − $20M))	$4.5M
Founders + employees	60% × $30M	$18M

The "2x" looks small on a term sheet. It costs founders $12M vs. Scenario A — 40% of their take at this exit price.

▸Scenario D — $15M down-side exit (1x non-participating)

Holder	Math	Payout
Series A	max($10M, 25% × $15M) → preference	$10M
Series Seed	max($2M, 15% × ($15M − $10M)) → preference	$2M
Founders + employees	what's left	$3M

The same cap table that returned $30M to founders at $50M returns $3M at $15M. Pref stacks compress like a spring at small exits and let go fully at big ones. Always model both — and the in-between.

Control: what the cap table doesn't tell you

The cap table tells you who owns the company. It does not tell you who runs it. Those are different, and the gap between them is one of the most important things to understand about VC deals.

At the time of a Series A, three things typically detach control from ownership:

Board composition

A typical post-Series A board is 5 seats: 2 common (founders), 2 preferred (lead investor + a co-investor or a jointly-agreed seat), 1 independent. The independent seat is usually selected by the preferred and common together, which gives the lead investor real influence over it. Founders can own 60% of the company and still not control the board. By Series B-C, founders are often the minority on their own board.

Protective provisions

The lead investor's preferred stock comes with consent rights — a list of major decisions the company can't take without the preferred class voting yes. The standard NVCA-style list includes:

• Selling the company or material assets
• Raising more money at any price (sometimes specifically: senior or pari-passu preferred)
• Increasing the option pool
• Paying dividends
• Taking on debt above a threshold
• Amending the charter or bylaws
• Changing the size of the board

A 25% Series A investor with full protective provisions has effective veto over every major decision the company makes. That's not 25% control. That's roughly 50/50 with the founders, even though the cap table says otherwise.

Drag-along

Most term sheets include a drag-along: if a defined threshold of preferred plus common votes to approve a sale, all shareholders are forced to consent. Common formulation: majority of preferred + majority of common (or majority of preferred + the board). This means a coalition of investors can force a sale founders don't want, by combining preferred votes with one founder's common vote, or with the board. Read every drag-along carefully — the exact thresholds determine who's actually deciding when the company gets sold.

▸Worked example — A 25% investor who effectively controls the company

Hypothetical. Investor Alpha leads a Series A, taking 25% at $40M post-money:

• Board: 1 of 5 seats (proportional)
• Protective provisions: standard NVCA list, including consent on raising new money, selling the company, and option pool increases
• Drag-along: majority preferred + majority common
• Class vote on Series B: required

Cap table reads 25%. Effective governance: Alpha can block any sale, any new financing, any option pool increase. Alpha doesn't run the day-to-day. But Alpha has structural veto over every direction the company can go.

This is the gap between the cap-table view of a company and the operating-control view. The terms determine which one you're buying. When you read a cap table, you're reading half the story — the term sheet that goes with it is the other half.

Anti-dilution

Anti-dilution kicks in when the company raises a down round: a round at a lower price per share than the previous round. Without protection, existing preferred holders see their percentage stake (and their pref math) get worse. Anti-dilution provisions adjust their conversion price to compensate.

There are three flavors. Two are normal. One is a red flag.

Broad-based weighted average

The industry standard. The conversion price for the existing preferred gets adjusted partially to reflect the new lower price, weighted by how many shares were issued at the down price relative to the total share count (including options, warrants, etc.). The math protects the investor proportionally without crushing common stockholders.

This is what you'll see in 95%+ of modern term sheets. Cooley's standard form term sheet uses it.

Narrow-based weighted average

Same idea, but with a smaller denominator (excludes options, warrants). Slightly more investor-favorable. Uncommon but not unusual.

Full ratchet

Conversion price drops to match the down round price. Dollar-for-dollar. The investor effectively keeps their full ownership stake regardless of how small the down round was. This crushes founders and the option pool.

If you see full ratchet in an early-stage term sheet, that's a red flag. Note it, raise it, expect to negotiate it out. The Holloway Guide and California Startup Lawyer both list it as a deal term founders should resist.

Pay-to-play and carve-outs

Anti-dilution provisions usually have carve-outs: share issuances that don't trigger the protection (option pool grants under a defined plan, M&A consideration, strategic partner shares). Some deals include pay-to-play: existing investors must participate in down rounds at their pro-rata or lose their anti-dilution protection (and sometimes get converted to common). Pay-to-play is rare in early-stage but worth knowing.

Common modeling traps

A few that show up in real diligence:

1. Confusing issued-and-outstanding with fully-diluted. Always ask which basis a percentage is on.
2. Treating the option pool as already granted. Most option pool numbers represent reserved shares, not issued options. Granted options dilute today; reserved-but-ungranted ones dilute when granted.
3. Stacked SAFEs creating accidental multi-x preferences. When several SAFEs convert at different caps in the same priced round, investors at the lower cap can end up with effective multi-x liquidation preferences. Mark Suster's fix: cap the conversion preference at 1x. Single paragraph in the SAFE side letter, saves a lot of grief.
4. Pre-money vs post-money SAFEs converting differently. Post-money SAFEs lock in a percentage; pre-money SAFEs are diluted by other SAFEs in the same round. Same cap, different outcomes. Run both.
5. Ignoring the option pool refresh at the next round. The option pool gets topped up at every priced round. Each refresh dilutes existing holders. If your model assumes the pool stays at 15% forever, you're underestimating dilution.
6. Modeling only one exit price. A deal that looks great at $200M might be terrible for founders at $30M. Model multiple scenarios and look at the slope of returns across them.

What you should be able to do after this

• Read a cap table and identify the issued-and-outstanding vs fully-diluted figures, the pref stack, and any non-standard provisions.
• Build a basic Excel model that walks dilution across rounds, including option pool refresh.
• Model an exit waterfall at multiple price points using the pref stack.
• Spot the red flags: full ratchet, multi-x prefs, uncapped participating, narrow-based anti-dilution, strict seniority.
• Tell a founder pitching you what their realistic ownership will look like at Series B if everything goes well, and if the next round is a down round.

That's a working baseline. The further reading below has the canonical deeper dives. Venture Deals (Feld and Mendelson), Secrets of Sand Hill Road (Kupor), and the Holloway equity guide. If you have a single Saturday, Venture Deals is the highest-ROI of the three.